Photo credit: Christian Wiediger/Unsplash

The results from this year’s Black Friday sales event suggest that the sway AI chatbots are having on ecommerce is only continuing to grow.

According to numbers from Salesforce, Black Friday purchases in Canada grew about 6% year-over-year to roughly $865 million USD, with AI and digital agents influencing around $173 million USD of that total. The CRM platform’s numbers showed a fivefold increase in AI-search traffic compared with Black Friday in 2024, while conversion rates on AI searches were four times higher than those from social media.

And data from Sensor Tower, a mobile-analytics and market-intelligence platform, shows that American shoppers increasingly relied on Amazon’s AI chatbot, Rufus, throughout November, peaking at 40% of app sessions on Black Friday, up from 30% at the start of the month. Sessions involving Rufus that resulted in a purchase grew disproportionately compared with non-AI sessions, Sensor Tower said, with a 91-point day-over-day spike in conversions on Nov. 28. A record $11.8 billion USD was spent during this year’s sales event in the U.S., according to Adobe Analytics.

The numbers suggest that AI-assisted sessions are becoming more effective at converting existing consumer intent, specifically during major sales events and on weekends when shoppers have more time to make deliberate purchases, according to Sensor Tower.

The battle for the ‘new gateway to commerce’

“We think this is going to be a massive disruption to how customers shop, and it’s going to be good for the consumer but tricky for retailers to navigate,” says Uwe Stueckmann, co-founder and partner of agentic AI consultancy Innovate Marketing. “And the way we’re thinking about it is sort of the battle for the new surface, the battle for the new gateway to commerce, if you will. The big question is, who’s going to win it?”

In recent months, Amazon launched Rufus in Canada, Walmart rolled out Sparky stateside and Google announced its own an agentic shopping service, while platforms like Shopify and Etsy have also introduced AI-powered initiatives. Stueckmann tells strategy that the pace of innovation points to a fundamental shift in how consumers will discover and evaluate products.

“I have no doubt that using an LLM agent to help you narrow down the product catalog in a way that considers all the things that are important to you is going to drive up conversion,” the former Loblaw executive says.

Preparing for a what’s next in retail

The shift poses a challenge for Canadian retailers that lack the scale and resources of global giants. While many brands have begun embedding AI into on-site search bars, Stueckmann says that approach only scratches the surface.

To help retailers prepare, Innovate Marketing has been working with smaller and mid-sized businesses to build branded AI agents within their owned ecommerce channels. That work, Stueckmann says, often starts with fundamentals like cleaning product data, ensuring reviews are accessible and plentiful and limiting login walls that prevent AI agents from indexing content.

“What we’re trying to get them to do is figure out how to create their own agents in their ecommerce environment, mostly to get ready for the launch of ChatGPT shopping and Gemini shopping in Canada,” Stueckmann says. “I think there’s got to be a bit of ‘Made in Canada’ that has to apply here to ensure that Canadian retailers have an opportunity to compete in the space and not be left behind. I think there’s a real risk around that.”

Personalization potential and ROI uncertainty

Personalization remains one of the clearest opportunities for retailers, particularly when AI is connected to first-party customer data across loyalty programs and owned channels, Elliott Marer says. According to the national leader for consumer and retail at KPMG Canada, a retailer could tap into past purchasing and browsing behaviour and implement AI-driven systems to surface more relevant product recommendations, like specific clothing that fits the fashion sense of a logged-in customer, or prioritizing discounted items for shoppers who consistently buy on sale.

“That’s where I think AI can really be powerful, because it’s coming up with a personalized product recommendation,” Marer says. “Whatever you might be looking for, it can then make suggestions based on your specific buying pattern.”

But during the relatively early stages of LLM strategy, Marer says organizations have tended to struggle to assess the return on investment for AI initiatives. In some cases, it might be prudent to invest early even if the ROI scope isn’t yet fully clear, just to stay competitive.

“Sometimes you have to be ready to change and invest in that change in order to continue your growth and sustain your business,” Marer says. “Because if you don’t, you might face a competitor that’s doing something more aggressively, and then you could become the forgotten brand.”

Consumer trust also remains a key constraint, particularly in Canada, where Marer says more than 80% of participants in a recent KPMG survey said they want to retain control over AI-assisted transactions rather than allowing systems to act autonomously.

“The people we surveyed want to have control of the AI transaction. They want to approve it all the way through,” Marer says, adding that regulatory and cultural differences also shape how far AI-enabled commerce can realistically go. “In Europe, there’s open banking where you can allow apps and and vendors right into your banking relationship to do certain things directly. And so I’m not sure we’re there yet.”

Tracking changes in consumer behaviour

While Canadians are known for being more cautious than their global counterparts when it comes to tech adoption, Olivier Blais sees the potential for that public position to change rather rapidly. The VP of AI at Montreal-based data-solutions company Moov AI says the growing use of virtual assistants reflects a broader shift in how consumers interact with technology.

“What we already know is that virtual assistants and the behaviours of chatting in natural language is wildly popular,” Blais tells strategy. “There are about 100 million daily users of ChatGPT that use the platform on average 25 times a day. That does not even include the millions using Google AI mode or any other generative AI platforms.”

That trend makes ecommerce support one of the most compelling uses for AI, he adds, with specialized shopping agents such as Amazon’s Rufus, Visa Intelligent Commerce, and Perplexity Shopping only becoming more popular.

“This surge in conversions and purchases supported by Rufus AI is very aligned with what we see in the market. It’s great to see such significant numbers on a big event like a Black Friday sale,” Blais says “We can even expect such agents to not only support consumers but even complete transactions soon.”